--- title: "ConocoPhillips Rod Pump Failure Dashboard" output: flexdashboard::flex_dashboard: orientation: columns vertical_layout: fill ---

Machine Learning Results

PCA 2 x PCA 1

ggplotly(empty_list[[1]]) %>% layout(legend = list(orientation = "h", x = 0.2, y = -0.2))

PCA 4 x PCA 1

ggplotly(empty_list[[3]]) %>% layout(legend = list(orientation = "h", x = 0.2, y = -0.2))

PCA 8 x PCA 2

ggplotly(empty_list[[13]]) %>% layout(legend = list(orientation = "h", x = 0.2, y = -0.2))

Multinomial Logistical Regression: Density Plot

ggplotly(ggplot(data2, aes(logit, fill=outcome))+geom_density(alpha=.3,) +
  geom_vline(xintercept=0,lty=2))

PAM Clustering

PAM Cluster: 2 x 2

ggplotly(ggpairs(pamclust5,columns=1:2 ,aes(color=as.factor(rodPumpScale$Failure_Type))))

3 x 3

ggplotly(ggpairs(pamclust5,columns=1:3 ,aes(color=as.factor(rodPumpScale$Failure_Type))))

4 x 4

ggplotly(ggpairs(pamclust5,columns=1:4 ,aes(color=as.factor(rodPumpScale$Failure_Type))))

5 x 5

ggplotly(ggpairs(pamclust5,columns=1:5 ,aes(color=as.factor(rodPumpScale$Failure_Type))))